Choosing feature selection methods for spatial modeling of soil fertility properties at the field scale Article Swipe

PDF

Related Concepts

Concordance correlation coefficient Covariate Feature selection Robustness (evolution) Random forest Computer science Best linear unbiased prediction Lasso (programming language) Cross-validation Decision tree Data mining Scale (ratio) Correlation coefficient Statistics Soil fertility Algorithm Artificial intelligence Mathematics Selection (genetic algorithm) Machine learning Soil science Environmental science Soil water World Wide Web Quantum mechanics Physics Gene Chemistry Biochemistry

Caner Ferhatoglu , Bradley A. Miller ·

YOU? · · 2022 · Open Access · · DOI: https://doi.org/10.1145/3557915.3565531 · OA: W4309651427

In this study, the effectiveness of six types of FS methods from four categories (filter, wrapper, embedded, and hybrid) were compared. These FS algorithms chose relevant covariates from a set of 1049 environmental covariates for predicting five soil fertility properties in ten fields, in combination with ten different ML algorithms. The resulting model performance was compared by three different metrics (R2 of 10-fold cross validation (CV), robustness ratio (RR; developed in this study), and independent validation with Lin's concordance correlation coefficient (IV-CCC)). Wrapper (BorutaShap) and embedded (Lasso-FS, Random forest-FS) methods with decision-tree based ML algorithms usually led to the optimal models.